METHOD FOR OPERATING AND CONTROLLING A WIND ENERGY PLANT

Description

[0001] The invention relates to a method for operating a wind turbine, in which a predictive controller calculates a trajectory for the wind turbine's power output based on a system state of the wind turbine. The wind turbine's power output is adjusted based on a trajectory value. The invention also relates to a control system for a wind turbine.

[0002] It is known to use a predictive controller for controlling a wind turbine (WO 2016 / 023560A1, WO 2016 / 023561A1). Predictive controllers are generally designed so that the calculated trajectory for the power output corresponds as closely as possible to the expected state of the wind turbine. A state in which the power to be delivered according to the trajectory deviates from the power available from the wind is undesirable and is avoided by appropriately calculating the trajectory.

[0003] The electrical energy generated by a wind turbine is typically fed into a grid operator's transmission grid. The goal is generally to operate the transmission grid at a constant grid frequency. However, a constant grid frequency is often impossible to maintain if there is a power balance error. If more power is fed into the transmission grid than consumers are using, the grid frequency increases. If consumers are using more power from the transmission grid than is being fed in, the grid frequency decreases. It is desirable for the wind turbines connected to the transmission grid to contribute to frequency stabilization.

[0004] In general, wind turbines have the option of reducing their power output. By appropriately adjusting the angle of attack of the rotor blades (pitch angle), the power consumption from the wind can be reduced, resulting in a corresponding reduction in power output. Conversely, it is generally not possible to increase the power output because wind turbines already absorb the optimal power from the wind during certain operating phases. This can be remedied by operating the wind turbine at reduced power during normal operation, so that more power can be drawn on demand at any time. However, this is associated with significant yield losses.

[0005] A short-term increase in power output is also possible by extracting kinetic energy from the rotor. However, this puts the rotor in an aerodynamically unfavorable state, so that the wind turbine can no longer even feed in the normal power immediately after the excess power has been extracted.

[0006] The invention is based on the object of presenting a method for operating a wind turbine and a control system so that the wind turbine can contribute to stabilizing the grid frequency without causing significant yield losses. Based on the cited prior art, the object is achieved with the features of the independent claims. Advantageous embodiments are specified in the subclaims.

[0007] In the method according to the invention for operating a wind turbine, a predictive controller calculates a trajectory for the power output of the wind turbine for a forecast period based on a system state of the wind turbine. The power output of the wind turbine is adjusted based on a trajectory value. In a first variant of the method, the trajectory comprises a section in which the trajectory values ​​correspond to an excess power compared to normal power and a section in which the trajectory values ​​correspond to an underpower compared to normal power. The excess power and the underpower add up to zero until the end of the forecast period.In a second variant of the method, the trajectory includes a section in which the trajectory values ​​correspond to an overperformance compared to normal performance and / or a section in which the trajectory values ​​correspond to an underperformance compared to normal performance. The overperformance and underperformance add up to zero by the end of the forecast period, whereby the addition of the overperformance and underperformance takes into account any performance surplus / deficit accumulated up to the beginning of the forecast period.

[0008] The invention recognizes that with an increasing proportion of wind turbines connected to a transmission grid, it becomes more difficult to keep the grid frequency stable. Previously, the base load was largely provided by synchronous generators. Synchronous generators have a self-stabilizing effect on the grid frequency by automatically feeding in more power when the grid frequency drops and less power when the grid frequency increases. Wind turbines do not have such a self-stabilizing effect on the grid. A corresponding stabilizing effect cannot be easily achieved by programming the wind turbine, because the energy absorbed from the wind cannot be increased at any time.

[0009] The invention proposes calculating a future-oriented trajectory for power output such that excess power and / or underpower is delivered during the forecast period. The deviation from normal power can be used to contribute to stabilizing the grid frequency, similar to a synchronous generator. A key element of the invention lies in not only allowing a deviation from normal power for the purpose of frequency stabilization, but also in calculating a path by which the wind turbine can be returned to a regular operating state. This is achieved by balancing the power balance by the end of the forecast period, with the excess power and underpower adding up to zero. Thus, favorable variations in output power for grid stabilization can be made, while simultaneously ensuring the balance of the power balance.

[0010] Through continuous, "flowing" application of the proposed method, it is possible to offer the continuous, slight variations in output power necessary for grid stabilization while simultaneously maintaining an averagely balanced power balance. This balanced power balance is important because otherwise, the under-power delivery would lead to yield losses and thus to revenue losses, while the over-power delivery would permanently lead either to an overload of the electrical system or to rotor stalling. Thus, the effect of the inventive method can also be understood as the targeted planning of the already existing fluctuations in the power delivered by wind turbines in such a way that they contribute to grid stabilization.

[0011] During normal operation, a wind turbine is typically controlled so that the electrical power generated corresponds to the power absorbed from the wind. The deviations between the power absorbed from the wind and the electrical power generated are generally kept as small as possible to prevent the rotor from accelerating or decelerating. The power output of a wind turbine during normal operation is referred to as the normal power.

[0012] Normal power is the power that the wind turbine would produce if it were not expected to produce excess or underpower. It results from the power provided by the wind and the boundary conditions that influence how optimally the wind turbine can convert this power into electrical energy. In normal operation, the wind turbine is operated in such a way that it converts the power absorbed from the wind directly and optimally into electrical energy. "Directly" means that there is no intermediate storage of the energy (e.g. as kinetic energy of the rotor, as thermal energy, as electrical energy in the intermediate circuit of the converter or in short-term electrical storage devices coupled to the turbine). An operating state in which energy is intermediately stored can generally only be continued for a limited time because the capacity of the available storage devices is limited.In contrast, the wind turbine can be operated for any length of time in normal operation.

[0013] To compare the future trajectory with a normal power output, a forecast of the normal power output over the forecast period is required. A necessary element of the forecast is the operating state of the wind turbine at the end of the forecast period, because this operating state is used to balance the power balance until the end of the forecast period. It is also possible to forecast the normal power output for one or more points in time within the forecast period.

[0014] In a simple case, the forecast of normal performance can consist of extrapolating the current normal performance for the duration of the forecast period. A forecast based on simple assumptions about changes expected during the forecast period is also possible. In one embodiment, the forecast for normal performance is derived from a normal trajectory, which is determined in a similar manner to the trajectory according to the invention. In contrast to the trajectory according to the invention, the normal trajectory can be calculated under the assumption that normal performance is provided throughout the forecast period.

[0015] A predictive controller estimates the operating state of a wind turbine for a future period based on the current system state. The variables that characterize the system state of the wind turbine include the operating parameters of the wind turbine and the operating parameters of the connecting grid.

[0016] The predictive controller can be designed to calculate a trajectory starting from a current time T 0 , which extends into the future over k points in time (sampling points). The time period (interval) between any two adjacent points in time T n and T n+1 should be chosen such that a trajectory can be calculated within a single calculation phase between T n and T n+1. A trajectory is therefore a sequence of trajectory values ​​extending from the current time T 0 to a future time TK. The time period between the points in time T 0 and TK is referred to as the prediction period.

[0017] The first trajectory value of a trajectory serves as a basis for setting the respective control variable of the wind turbine in the following interval.

[0018] At the beginning of each new interval, data on the actual system state of the wind turbine in the previous interval is available. The predictive controller creates another trajectory in the new interval, shifted backward by one interval. This new trajectory serves as the basis for adjusting the wind turbine's control variable for the subsequent interval. The predictive controller can be designed, in particular, as a model predictive control (MPC, Receding Horizon Control).

[0019] The manipulated variables to which the trajectory can refer include the torque applied to the rotor by the electrical system, the angle of attack of one or more rotor blades (pitch angle), and the angle at which the turbine is aligned with the wind (azimuth angle). Variables derived from these variables, such as the power output of the wind turbine, also belong to the manipulated variables. A trajectory can refer to one manipulated variable, to several manipulated variables, or to the course of states (e.g., the rotational speed or the tower head deflection) of the wind turbine. Within the scope of the present invention, the trajectory value of such a trajectory related to the power output is considered.

[0020] A trajectory value is one of a sequence of trajectory values ​​that together form the trajectory. If a wind turbine's manipulated variable is adjusted based on the trajectory value, a setpoint and / or a default value for the manipulated variable is derived from the trajectory value. A design in which the predictive controller provides both a setpoint and a feedforward value is known as a two-degree-of-freedom structure.

[0021] A preset value can be set directly in the corresponding actuator of the wind turbine or superimposed on an output value of a controller. A setpoint can be specified to a subordinate controller, which controls the manipulated variable taking into account the setpoint and measured values. The setpoint or preset value can be derived from a single trajectory value. It is also possible for the setpoint or preset value to be derived from several trajectory values ​​or from an interpolation between several trajectory values.

[0022] The trajectory may include a first section in which a deviation in the grid frequency is counteracted by the deviation from the normal power. The trajectory may include a second section in which a deviation that developed in the first section is compensated. When calculating the trajectory in the second section, it can be taken into account that the wind turbine may have entered an aerodynamically unfavorable state due to the power output in the first section.

[0023] It is advantageous if the second section follows the first section in time. This makes it possible to quickly counteract a frequency change. The fact that the deviation from normal power in the second section no longer necessarily counteracts the frequency deviation can be accepted. After all, the grid must be able to stabilize itself with a certain lead time, regardless of the contribution from the wind turbines. The contribution from the wind turbines can also decrease, for example, due to a decrease in wind speed.

[0024] When calculating the trajectory, a grid frequency forecast can be taken into account. The grid frequency forecast can be provided by the transmission system operator. It is also possible for the grid frequency forecast to be determined from data available locally at the wind turbine or wind farm. The grid frequency forecast can cover the entire forecast period or part of it. Preferably, the grid frequency forecast covers a period immediately following the current time.

[0025] A forecast for the power to be delivered can be derived from the grid frequency forecast. Assuming that a normal trajectory would result from a constant grid frequency over the forecast period, deviations in the grid frequency can be expressed as overpower or underpower. To simulate the behavior of a synchronous generator, the trajectory can include a section in which the overpower or underpower is proportional to the change in frequency. Additionally or alternatively, the trajectory can include a section in which the overpower or underpower is proportional to the frequency deviation.

[0026] The excess power or underpower derived from the grid frequency forecast indicates which excess power or underpower would be desirable over the forecast period to stabilize the grid (desired power forecast). It cannot generally be assumed that the desired power forecast corresponds to what the wind turbine can actually provide. The trajectory according to the invention is calculated such that the desired excess power or underpower is provided to an extent that corresponds to the actual capabilities of the wind turbine. The desired power forecast can be derived from the grid frequency forecast within the method according to the invention. It is also possible that the desired power forecast is already available and can be used as an input variable in the method according to the invention.

[0027] When determining the trajectory according to the invention, one or more of the following boundary conditions can be considered. The trajectory can be determined such that the excess power and the underpower add up to zero by the end of the forecast period. This is particularly advantageous if the wind turbine is in a balanced operating state at the beginning of the forecast period, so that there is no excess power or deficit that has built up in the past. Furthermore, the excess power or deficit resulting from the changed operating conditions (e.g., the aerodynamic power coefficient or loss factors) due to the deviation from the normal power can be taken into account.If there is a power surplus / deficit at the beginning of the forecast period (for example, because wind conditions have changed unexpectedly shortly beforehand), the trajectory can be determined such that the initial power surplus / deficit, the overpower during the forecast period, and the underpower during the forecast period add up to zero by the end of the forecast period. If there is a power surplus or a power deficit at the beginning of the forecast period, it may occur that only overpower or only underpower is delivered throughout the entire forecast period. If the power balance is balanced at the beginning of the forecast period, the trajectory defines at least one phase of overpower and at least one phase of underpower within the forecast period.

[0028] Since the power balance is also reflected in the energy state of a turbine, determining a balanced balance, in which the overperformance and underperformance add up to zero by the end of the forecast period, does not necessarily require a comparison with the normal power or a balancing of overperformance and underperformance. A balanced balance can also be determined, for example, by the state of the turbine at the end of the trajectory under consideration. The trajectory is balanced when all energetic end states have the same value as if a normal trajectory had been followed. In particular, the power output, blade angle, and speed can correspond to the normal values.

[0029] According to a further boundary condition, the trajectory according to the invention can coincide with the normal trajectory at the end of the prediction period. The trajectory can be determined such that the difference between the trajectory and the normal trajectory over the entire prediction period is smaller than a predefined first threshold. The trajectory can be determined such that the gradient of the trajectory is not greater than a predefined second threshold. The trajectory can be determined such that the deviation between the trajectory and the desired performance forecast can increase with the increasing duration of the prediction period. For example, an exponential weighting that decreases over the prediction period can be provided, which allows for larger deviations between the trajectory and the desired performance forecast as the duration of the prediction period increases.

[0030] The desired performance forecast taken into account when determining the trajectory can extend over the entire forecast period. The trajectory can be derived from the desired performance forecast using a best-fit curve. The best-fit curve can be designed so that it runs partially above the desired performance forecast and partially below the desired performance forecast. In particular, the best-fit curve can be dimensioned such that the area enclosed above the best-fit curve with the desired performance forecast is exactly the same size as the area enclosed below the best-fit curve with the desired performance forecast. The trajectory can be determined as the difference between the performance forecast and the best-fit curve. In one embodiment, the best-fit curve is designed as a best-fit straight line.

[0031] A compensation curve in the form of a counter-curve is also possible. The counter-curve can be generated by mirroring the desired performance forecast with the normal trajectory and running in the opposite direction. The trajectory can be formed as a linear superposition of the desired performance forecast and the counter-curve, so that the overperformance and underperformance balance each other out within the forecast period.

[0032] In another embodiment, the trajectory is determined based on a desired performance forecast that extends only over part of the forecast period. For example, the performance forecast can define a pattern of overperformance or underperformance for a portion of the forecast period. The trajectory can coincide with the desired performance forecast in this portion. In the remaining portion of the forecast period, the trajectory can be continued such that the overperformance and underperformance balance each other out by the end of the forecast period, and the trajectory coincides with the normal trajectory at the end of the transfer horizon.

[0033] It is also possible for the desired performance forecast to define a gradient at which the overperformance or underperformance increases from the current point in time. The trajectory can initially adopt this gradient. In a later portion of the forecast period, the trajectory can be continued in such a way that the overperformance and underperformance balance out by the end of the forecast period, and that the trajectory coincides with the normal trajectory at the end of the forecast period.

[0034] In addition to or as an alternative to the desired power forecast, a forecast of the behavior of an inverter can be taken into account when determining the trajectory. The inverter can be designed to independently adjust the power output on the grid side to the needs of the grid. For example, the inverter can be designed as a virtual synchronous machine that adjusts the power output depending on the grid frequency and / or a rate of change of the grid frequency. The inverter can be a full-scale converter through which the entire electrical power generated by the wind turbine is transmitted. In particular, the intermediate circuit voltage of the inverter can be taken into account when determining the trajectory.

[0035] The invention also relates to a controller for a wind turbine. The controller comprises a predictive controller designed to calculate a trajectory for a manipulated variable of the wind turbine for a forecast period based on a system state of the wind turbine. The controller is designed to generate a control signal to adjust the manipulated variable of the wind turbine based on a trajectory value. In a first variant, the trajectory comprises a section in which the trajectory values ​​correspond to an excess power compared to normal power and a section in which the trajectory values ​​correspond to an underpower compared to normal power. The excess power and the underpower add up to zero until the end of the forecast period.In a second variant, the trajectory comprises a section in which the trajectory values ​​correspond to an excess power compared to normal power and / or a section in which the trajectory values ​​correspond to an underpower compared to normal power. The excess power and the underpower add up to zero by the end of the forecast period, whereby the addition of the excess power and the underpower takes into account any power surplus / deficit accrued up to the beginning of the forecast period. The invention also relates to a wind turbine that comprises such a control system and / or that is operated under the control of such a control system.

[0036] The control system can be further developed with additional features that are described in connection with the method according to the invention. The method can be further developed with additional features that are described in connection with the control system according to the invention.

[0037] The invention is described below by way of example with reference to advantageous embodiments in the accompanying drawings. They show: Fig. 1: a schematic representation of a wind turbine according to the invention; Fig. 2: a control system of a wind turbine in a first embodiment of the invention; Fig. 3: the view according to Fig. 2 in another embodiment of the invention; Fig. 4: a normal trajectory of the wind turbine; Fig. 5: a schematic representation of steps for determining a trajectory according to the invention; Figs. 6 to 10: the determination of the trajectory in alternative embodiments of the invention.

[0038] In a Fig. 1 In the wind turbine shown, a nacelle 14 is rotatably mounted on a tower 15. The nacelle carries a rotor 16, which is set in rotation by the wind. The rotor is connected via a gear 17 to a generator 18, which generates electrical energy when the rotor 16 rotates. The electrical energy is converted into a Fig. 1 not shown connection network.

[0039] A controller 20 provides specifications for the operation of the wind turbine. These specifications relate to the wind turbine's control variables, which include, in particular, the pitch angle of the rotor blades and the settings of the generator 18 and the converter 19. The generator 18 and the converter 19 together form an electrical system that counteracts the rotor's torque.

[0040] The control 20 comprises according to Fig. 2 a state estimator 21, which determines the system state of the wind turbine based on measured values ​​22. Some parameters characterizing the system state of the wind turbine are accessible to direct measurement. Other parameters are estimated based on the available measured values.

[0041] The information about the system state is fed to a predictive controller 23, which, based on the current system state, calculates a future trajectory for the wind turbine's control variables. The trajectory specifies a course of the control variables that delivers an optimal result based on a state model of the wind turbine stored in the predictive controller 23 and taking into account a specific quality criterion.

[0042] A trajectory value 24 is fed to the input of a subordinate controller 25 as a setpoint. The subordinate controller 25 processes the trajectory value 24 and the measured values ​​22 and uses them to determine a target value 26 for the pitch angle of the rotor blades and the settings of the electrical system 18, 19. The relevant parameters of the wind turbine are adjusted via actuators according to the target values ​​26 of the subordinate controller 25.

[0043] In the alternative embodiment according to Fig. 3 The trajectory value 24 is superimposed on the output signal of the subordinate controller 25 by means of a feedforward control. In another alternative, the predictive controller provides both a setpoint and a feedforward control value (two-degree-of-freedom structure).

[0044] In Fig. 4 a normal trajectory 27 for the electrical power generated by the wind turbine is plotted over time t, which has been calculated by the predictive controller 23. Up to a time T 0 , measured values ​​are available for the actual curve 28 of the electrical power of the wind turbine. At time T 0 , a calculation phase 29 begins which extends up to a time T 1 . In a first section of the calculation phase 29, the state estimator 21 estimates a system state vector, which is fed to the predictive controller 21. In the second section of the calculation phase 29, the predictive controller 23 calculates the normal trajectory 27 based on the system state vector, which extends from the time T 1 to a future time TK. The time period between the times T 1 and TK is referred to as the prediction period.When calculating the normal trajectory 27, a wind forecast is taken into account that extends over the forecast period.

[0045] The normal trajectory 27 is composed of sampling points 35, with the temporal separation between two adjacent sampling points 35 being at least as long as the length of the calculation phase 29. The trajectory 27 is linearly interpolated between each two adjacent sampling points 35. In total, the trajectory 27 can extend, for example, over 200 intervals into the future.

[0046] At time T 1 , the calculation of trajectory 27, which extends from time T 1 to time TK, is completed. The subordinate controller 25 regulates between times T 1 and T 2 based on the course of trajectory 27 between times T 1 and T 2 .

[0047] The further course of trajectory 27 after time T 2 has no influence on the wind turbine's control variables. At time T 1, measured values ​​from the interval between T 0 and T 1 are available. Based on these measured values, a new trajectory is calculated in the interval between T 1 and T 2, which specifies the course of the control variable between times T 2 and T 3.

[0048] Using the normal trajectory 27, the predictive controller 23 calculates an optimal power output curve from the perspective of the wind turbine during the forecast period. An optimal curve in this sense means, in particular, that the electrical power generation is aligned as closely as possible with the power absorbed from the wind.

[0049] In the method according to the invention, a trajectory deviating from the normal trajectory is calculated, in which overpower and / or underpower is delivered in phases within the transmission horizon deviating from the normal trajectory and taking into account a balanced power balance.

[0050] In Fig. 5 The horizontal axis coincides with the power that the wind turbine would deliver within the transmission horizon according to the normal trajectory 27. For example, the predictive controller 23 receives a desired power forecast 30 for the forecast period from the transmission grid, see Fig. 5A The desired power forecast 23 deviates upwards from the normal trajectory 27, so that the wind turbine would have to deliver excess power throughout the entire forecast period to meet the desired power forecast 30. This is not possible because the wind turbine already converts all the energy absorbed from the wind into electrical power according to the normal trajectory 27.

[0051] The aim of the method according to the invention is to determine a trajectory 31 in which the desired performance forecast 30 is at least partially taken into account. In the embodiment of the Fig. 5 To do this, a best-fit line 32 is first drawn through the desired performance forecast 30, so that the areas enclosed above and below the best-fit line 32 with the desired performance forecast 30 are of equal size. When determining the compensation, the reduced performance resulting from the aerodynamically less efficient operation when deviating from the normal trajectory can also be taken into account.

[0052] The Fig. 5C The trajectory 31 according to the invention shown is determined by adding the difference between the desired performance forecast 30 and the best-fit line 32 to the normal trajectory 27 for each point in time within the forecast period. In sections in which the desired performance forecast 30 is greater than the best-fit line 32, the trajectory 31 defines an overperformance 33 compared to the normal trajectory 27. In sections in which the desired performance forecast 30 is smaller than the best-fit line 32, the trajectory 31 defines an underperformance 34 compared to the normal trajectory 27.

[0053] The amount of excess power 33 during the forecast period is the same as the amount of underpower 34. Thus, the electrical power generated according to the trajectory 31 during the forecast period corresponds to the power planned according to the normal trajectory 27.

[0054] In the alternative embodiment according to Fig. 6 only the first section between times T 1 and TM of the desired power forecast 30 is taken into account when determining trajectory 31 by transferring this section identically to trajectory 31. The second section of the transmission period between times TM and TK is used to return the wind turbine to a normal operating state. For this purpose, trajectory 31 is completed with a curve section 36 that smoothly connects to the first section, compensates for the excess power 33 with a corresponding amount of underpower 34, and at the end of the transmission horizon at time TK again coincides with and smoothly transitions into the normal trajectory 27. A curve shape can be defined that satisfies these criteria and only needs to be scaled in order to form a connection to the first section of trajectory 31.

[0055] In Fig. 7 an embodiment is shown in which a counter curve 37 is formed for the desired performance forecast 30, the shape of which corresponds to the desired performance forecast 30, see Fig. 7A . The counter curve 37 is mirrored in relation to the desired performance forecast 30 on the normal trajectory 27 and runs in the opposite direction from time TK to time T 1 . By linearly superimposing the desired performance curve 30 and the counter curve 37, the Fig. 7B The trajectory 31 shown is formed, which also satisfies the aforementioned criteria. The amount of overperformance 33 is thus equal to the amount of underperformance 34, and at time TK, trajectory 31 transitions smoothly into the normal trajectory 27.

[0056] In the embodiment according to Fig. 8 The desired power forecast 30 is limited to the information that the desired power increases at a certain gradient at time T1. The trajectory 31 takes up this gradient at time T1 and, after an initial phase of increase, moves away from this gradient in order to return the wind turbine to normal operation within the forecast period. Fig. 8 The trajectory 31 shown satisfies the requirements that the overperformance 33 is equal to the underperformance 34 and that the trajectory 31 transitions smoothly into the normal trajectory 27 at time TK.

[0057] In Fig. 9 Only the desired performance forecast 30 and the normal trajectory 27 are shown. The trajectory 31, not shown, is determined as an optimization task taking the following boundary conditions into account. The sum of overperformance and underperformance during the forecast period is zero. The deviation between the trajectory 31 and the normal trajectory at the end of the forecast period is zero. The deviation between the trajectory 31 and the normal trajectory 27 is not greater than a predefined threshold. The gradient of the trajectory 31 is never greater than a predefined second threshold. The deviation between the trajectory 31 and the desired performance forecast 30 is as small as possible, with only small deviations being permitted at the beginning of the forecast period and larger deviations being permitted as the forecast period progresses.For example, an exponential weighting can be provided, allowing increasingly larger deviations from the desired performance forecast. Solving this optimization problem leads to the desired trajectory 31.

[0058] In the previous embodiments, the determination of the trajectory 31 has been described under the assumption that the actual operation of the wind turbine corresponded to the normal trajectory 27 up to the beginning of the forecast period, i.e. that no surplus / deficit in the power output had accumulated up to the beginning of the forecast period.

[0059] In Fig. 10 An example is shown in which the wind turbine has provided a power surplus 38 up to time T1, which is Fig. 10is shown as a rectangle. In such a case, the trajectory 31 is determined such that the sum of the power surplus / deficit 38 existing at the beginning of the forecast period, the excess power 33 during the forecast period, and the underpower 34 during the forecast period add up to zero by the end of the forecast period. In addition to the aerodynamic and kinetic state of the wind turbine, the power surplus / deficit can also take into account the state of charge of the converter intermediate circuit and / or other short-term electrical storage devices coupled to the turbine.

Claims

1. Method for operating a wind turbine, in which a predictive controller (23) uses a system state of the wind turbine as a basis for calculating a trajectory (31) for the power output of the wind turbine for a forecast period and in which the power output of the wind turbine is adjusted on the basis of a trajectory value (35), a. wherein the trajectory (31) comprises a section in which the trajectory values (35) correspond to an overpower (33) relative to normal power, and comprises a section in which the trajectory values (35) correspond to an underpower (34) relative to normal power, characterized in that the overpower (33) and the underpower (34) add up to zero by the end (TK) of the forecast period, or b. wherein the trajectory (31) comprises a section in which the trajectory values (35) correspond to an overpower (33) relative to normal power, and / or comprises a section in which the trajectory values (35) correspond to an underpower (34) relative to normal power, characterized in that the overpower (33) and the underpower (34) add up to zero by the end (TK) of the forecast period, wherein a power surplus / deficit accrued up to the beginning of the forecast period is taken into account when overpower (33) and underpower (34) are being added up.

2. Method according to Claim 1, characterized in that a prediction about the normal power is made during the forecast period.

3. Method according to Claim 1 or 2, characterized in that the normal power is derived from a normal trajectory (27), wherein the normal trajectory (27) is calculated under the assumption that normal power is provided during the forecast period.

4. Method according to one of Claims 1 to 3, characterized in that the trajectory (31) comprises a first section in which a deviation of the mains frequency is counteracted by the deviation from the normal power.

5. Method according to Claim 4, characterized in that the trajectory (31) comprises a second section in which a deviation from the normal power built up in the first section is compensated for.

6. Method according to one of Claims 1 to 5, characterized in that the trajectory (31) is derived from a desired power prediction (30).

7. Method according to one of Claims 1 to 6, characterized in that a reduced power resulting from the aerodynamically less efficient operation when deviating from the normal trajectory (27) is taken into account when overpower (33) and underpower (34) are being added up.

8. Method according to one of Claims 1 to 7, characterized in that the trajectory (31) at the end (TK) of the forecast period corresponds to the normal trajectory (27).

9. Method according to one of Claims 1 to 8, characterized in that a prediction about the behaviour of a converter is taken into account when the trajectory (31) is being determined.

10. Method according to Claim 9, characterized in that the trajectory (31) is calculated taking into account a DC link voltage of the converter.

11. Control system for a wind turbine, having a predictive controller (23) which uses a system state of the wind turbine as a basis for calculating a trajectory (27) for a manipulated variable of the wind turbine for a forecast period, wherein the control system is designed to generate a control signal (26) in order to adjust the manipulated variable of the wind turbine on the basis of a trajectory value (35), a. wherein the trajectory (31) comprises a section in which the trajectory values (35) correspond to an overpower (33) relative to normal power, and comprises a section in which the trajectory values (35) correspond to an underpower (34) relative to normal power, characterized in that the overpower (33) and the underpower (34) add up to zero by the end (TK) of the forecast period, or b. wherein the trajectory (31) comprises a section in which the trajectory values (35) correspond to an overpower (33) relative to normal power, and / or comprises a section in which the trajectory values (35) correspond to an underpower (34) relative to normal power, characterized in that the overpower (33) and the underpower (34) add up to zero by the end (TK) of the forecast period, wherein a power surplus / deficit accrued up to the beginning of the forecast period is taken into account when overpower (33) and underpower (34) are being added up.